NeuroCAD

Intelligence to Structure.

OVERVIEW

NeuroCAD is a next-generation CAD kernel built on implicit geometry (Signed Distance Fields) with B-Rep hybrid capabilities. It represents a fundamental rethinking of CAD architecture — where truth is the construction graph and field, not the mesh.

With 42 modules across 5 tiers, NeuroCAD covers parametric modeling, simulation (FEA/CFD/CAM), generative design, real-time collaboration, and AI-assisted design through the NACP protocol.

FIVE PILLARS

AI-Native Protocol (NACP)

Intent-based design commands. Natural language → parametric features.

Hybrid Geometry

B-Rep + F-Rep/SDF seamlessly mixed. Traditional and generative in one kernel.

GPU-Accelerated Solver

WebGPU compute shaders. 10x faster constraint solving.

Semantic Feature Recognition

ML-powered detection of holes, pockets, slots, threads from STEP imports.

Live Collaboration (CRDT)

Google Docs for 3D CAD. Automerge-based, zero conflicts.

TECHNOLOGY
RustKernel
WebGPURendering
13Rust crates
CRDTCollaboration

Ready to design with intelligence?